Persistence of Subclinical Intramammary Pathogens in Goats Throughout Lactation

Persistence of Subclinical Intramammary Pathogens in Goats Throughout Lactation A. CONTRERAS, J. C. CORRALES, A. SANCHEZ, and D. SIERRA Enfermedades Infecciosas, Departamento de Patologia Animal, Facultad de Veterinaria, Universidad de Murcia, Campus Universitario de Espinardo, 30071 Murcia, Spain Received October 15, 1996. Accepted April 29, 1997. ABSTRACT The goal of this study was to determine the persistence of caprine intramammary pathogens throughout lactation and to detect the bias in diagnoses when a single milk sample was used. We studied 131 goats throughout 7 mo of lactation. Goats were sampled monthly, and 1834 milk samples were bacteriologically analyzed. One hundred sixty-eight pathogens were isolated: 82.5% were micrococci, 9.5% were Gram-negative bacilli, and 8% were corynebacteria. An intramammary infection (IMI) was considered a true, persistent IMI when the same pathogen was isolated two or more times consecutively from the same half of the udder. One hundred one samples were considered to be truly positive, which produced persistent IMI caused by nine different species (eight Staphylococcus spp. and one Pseudomonas sp.). Statistical relationships were found between staphylococci and true-positive diagnosis and between corynebacteria and false-positive diagnosis. No relationship involving Gram-negative bacilli was detected. A single milk sample had a positive predictive value (60%), high sensitivity (96.2%), high specificity (96.1%), and highly negative predictive value (99.8%). ( Key words: persistence, subclinical mastitis, staphylococci, goat) Abbreviation key: CNS = coagulase-negative staphylococci, GNB = Gram-negative bacilli. INTRODUCTION Mastitis is one of the most costly diseases in the dairy industry, and, although much information is available concerning mastitis in cows, few studies deal with mastitis in goats. Researchers studying subclinical mastitis in goats agree that IMI caused by coagulase-negative staphylococci ( CNS) are the most prevalent (4, 8, 14, 15, 16, 21). Despite the high prevalence of IMI caused by CNS, CNS are considered to be minor pathogens. However, IMI caused by CNS are associated with clinical mastitis, changes in milk composition, and reduced milk yield (12, 13); IMI caused by CNS are also capable of persisting throughout lactation and the dry period (21). Use of a single milk sample for bacteriological analyses to detect IMI in dairy cows might contribute to misdiagnosis (18). This study was designed to determine the monthly persistence of caprine IMI throughout lactation and to detect the bias in diagnoses when a single milk sample was used for bacteriological analysis. Goats MATERIALS AND METHODS Four goat herds from the Asociacion Española de Criadores de la Cabra Murciano-Granadina (ACRIMUR-Jumilla, Murcia, Spain) were used, including a total of 447 lactating goats. Thirty percent of the goats were selected within 45 d after parturition by systematic random sampling (24) at the milking parlor. Milk samples were collected monthly during 225 d of lactation. Of the 134 goats selected, 3 were excluded because of clinical mastitis; thus, a total of 1834 milk samples was analyzed from 131 lactating goats of varying ages. These herds had been included in a program for the control of subclinical mastitis and were free of caprine mycoplasmosis. Does were machine-milked; to control mastitis, a teat dip with an iodine solution was applied after milking, and antibiotic dry therapy was practiced. In the previous lactation, these herds had a low incidence of subclinical mastitis (prevalence of IMI per gland ranged from 10 to 15%) (22). Sample Collection Before the morning milking, teats were carefully cleaned with 96% ethanol, and the first three streams of foremilk were discarded. Ten milliliters of milk 1997 J Dairy Sci 80:2815 2819 2815

2816 CONTRERAS ET AL. were collected aseptically from each gland. Samples were kept at 4 C for 2 to 4 h until bacteriological analysis. Bacteriological Procedure Twenty microliters of each sample were plated on blood agar plates (5% washed sheep erythrocytes). The plates were incubated aerobically at 37 C and examined at 24, 48, 72, and 168 h. An IMI was determined at 250 cfu/ml. Cultures with five or more identical colonies were considered positive for an IMI. Bacteria were identified according to the recommendations of the National Mastitis Council (11). Identification of staphylococci was made using commercial micromethods (API STAPH; BioMèrieux, Lyon, France). Analysis When the same pathogen was isolated two or more times from the same half of the udder, it was considered a true-positive diagnosis of IMI ( 1 ) (Table 2). When the same pathogen was isolated only once from the same half of the udder, the diagnosis was considered a false-positive diagnosis of IMI. Validity parameters of a single bacteriological analysis (sensitivity, specificity, and predictive values) versus multiple analyses were calculated according to the recommendations of Thrusfield (24). Confidence intervals (95%) for these associated measures were calculated using Epi Info (6). To demonstrate the association between different groups of pathogens and true IMI, we performed a chi-square distribution test using 2 2 contingency tables. Odds ratios and Cornfield 95% confidence limits were used to explain the type of association calculated using Epi Info (6). RESULTS Among the 1834 samples analyzed, bacteria were recovered from 168 samples (9% prevalence). Micrococci was the most frequently isolated group of pathogens (82.5%). Other pathogens isolated were Gram-negative bacilli ( GNB) (9.5%) and corynebacteria (8%). No streptococci or bacilli were identified. Of the 139 Micrococcaceae isolated, 93.5% were identified by the API STAPH system. One isolate was Micrococcus sp. Of the 130 staphylococci identified, Staphylococcus chromogenes, Staphylococcus xylosus, and Staphylococcus caprae were the pathogens most frequently isolated (Table 1); 16 isolates were GNB (4 coliforms and 12 Pseudomonas spp.), and 13 isolates were Corynebacterium spp., but specific identification was not realized. TABLE 1. Staphylococci identified from 130 isolates. Species (no.) (%) Staphylococcus xylosus 31 23.8 Staphlococcus chromogenes 30 23 Staphylococcus caprae 14 10.8 Staphylococcus capitis 12 9.2 Staphylococcus aureus 10 7.7 Staphylococcus warneri 10 7.7 Staphylococcus simulans 7 5.4 Staphylococcus epidermidis 6 4.6 Staphylococcus lentus 5 4.6 Staphylococus hominis 3 2.4 Staphylococcus sciuri 1 0.7 Staphylococcus haemolyticus 1 0.7 Total 130 100 Of the 262 udder halves studied, 17 IMI were detected (6.5% prevalence); these IMI were caused by nine separate pathogens, which yielded a total of 101 true-positive isolates (Table 2). Eight of the pathogens were staphylococci, and one pathogen was a Pseudomonas sp.; the goat infected with the Pseudomonas sp. developed clinical mastitis later in lactation and was removed from the herd. No true persistent IMI caused by corynebacteria were detected. The most persistent staphylococci isolates were Staph. chromogenes and Staph. xylosus; but Staph. caprae, Staphylococcus capitis, Staphylococcus warneri, Staphylococcus lentus, Staphylococcus aureus, and Staphylococcus epidermidis were also detected. Duration of IMI ranged from two to seven consecutive samplings. Ten halves were infected throughout the study (Table 2). On four occasions, bacteria were not isolated between persistent IMI periods; these results were considered false negatives. A total of 67 false positives was detected: 45 of these were caused by staphylococci, 13 were caused by corynebacteria, and 9 were caused by GNB. For a single bacteriological analysis versus multiple analyses, measures of association and 95% confidence intervals, respectively, were sensitivity, 96.2% and 90.0 to 98.8%; specificity, 96.1% and 95.1 to 97.0%; predictive positive value, 60.1% and 52.3 to 67.5%; and predictive negative value, 99.8% and 99.2 to 99.9%. When groups of pathogens were considered (Table 3), the predictive positive value increased to 67.6% (59.1 to 75.2%) for staphylococci and decreased to 43.8% (20.8 to 69.4%) for GNB. Because all corynebacteria yielded false-negative results, those validity parameters could not be calculated for this group. Statistical associations were found between staphylococci and true-positive diagnoses ( P < 0.005)

TABLE 2. Number and duration of persistent 1 IMI. INTRAMAMMARY PATHOGENS IN GOATS 2817 Persistent Sampling detected 2 Species IMI (no.) Period Isolates (no.) (%) Staphylococcus chromogenes 5 28 27.7 7 3 1 7 6 1 6 2 6 7 Staphylococcus xylosus 4 7 3 1 7 21 20.8 6 2 7 2 6 7 Staphylococcus caprae 2 7 3 1 7 13 12.8 Staphylococcus capitis 2 6 2 7 11 10.9 5 3 7 Staphylococcus warneri 1 7 6.9 Staphylococcus lentus 1 7 6.9 Pseudomonas sp. 1 7 6.9 Staphylococcus aureus 1 5 3 3 7 4 3.9 Staphylococcus epidermidis 1 3 3 5 3 2.9 Total 17 101 100 1The same pathogen was isolated two or more times from the same half of the udder. 2The number of samplings in which the pathogens were consecutively isolated and the duration of this period. 3Bacteria were not detected on one occasion. and between corynebacteria and false-positive diagnoses ( P < 0.005). No relationship involving GNB was detected. DISCUSSION Similar to the results obtained by other researchers (3, 4, 5, 6, 7, 8, 9, 10, 14, 17, 21), the present study showed a high percentage of IMI caused by micrococci and a low percentage of IMI caused by other pathogens. The API STAPH system identified 93.5% of the micrococci isolated in this study. Previous studies have also reported high percentages of identified staphylococci in goats. Kalogridou-Vassiliadou (15) and Contreras et al. ( 4 ) reported that 96.9 and 95.9%, respectively, of IMI caused by micrococci were identified using the API STAPH system; Deinhofer and Pernthaner ( 8 ) identified 84% of the infections using the ATB 32 STAPH system (BioMèrieux). These percentages were higher than the 70% found by Honkanen-Buzalski (13) in cows and the 82.4% found by Bergonier et al. ( 2 ) in sheep using the API STAPH method. The literature reports a wide range of unidentified staphylococci in other ruminant species, varying from 13 to 20% in sheep and from 42 to 60% in cows (2). Data from this and other studies (4, 8, 15) indicate that commercial micromethods are a quick and easy way to identify staphylococci that have been obtained from the caprine mammary gland. In this study, Staphylococcus spp. and Pseudomonas aeruginosa were the only bacteria detected in persistent IMI. Eight species of Staphylococcus spp. TABLE 3. True- and false-positive diagnoses for each group of pathogens and statistical results to estimate the capability of each group to produce true-positive diagnoses. Validity parameters with 95% confidence intervals [sensitivity, specificity, predictive positive value (PPV), and predictive negative value (PNV)] are provided for staphylococci and Gram-negative bacilli (GNB). Staphylococci GNB Corynebacteria (n = 139) (n = 16) (n = 13) True positives 94 7 0 False positives 45 9 13 False negatives 4 0 0 True negatives 1662 1666 1666 x 2 17.6 1.29 18.61 P <0.001 >0.05 <0.001 Odds ratio 6.57 0.48 0.00 (2.4 18.4) 1 (0.1 1.5) (0.0 0.2) Sensitivity, % 95.9 100... (89.3 98.7) (56.1 100)... Specificity, % 97.4 99.5... (96.5 98) (98.9 99.7)... PPV, % 67.6 43.8... (59.1 75.2) (20.8 69.4)... PNV, % 99.8 100... (99.3 99.9) (99.7 100)... 1Numbers in parentheses are 95% confidence intervals.

2818 CONTRERAS ET AL. were detected during 2 or more consecutive mo. Staphylococcus chromogenes was the most frequently isolated species and has been shown to be the predominant CNS species isolated from the bovine streak canal (12). Staphylococcus xylosus, the second most frequently isolated species in this study, is the predominant isolate from the teat skin and other external bovine surfaces (12). Our data agree with that determination because Staph. xylosus was the most prevalent in total IMI, but not in persistent IMI. Little information is available concerning Staph. caprae, the third most frequently isolated species in this study. This species is capable of persisting throughout the dry period (21) and was one of the most common species isolated from the caprine mammary gland in several studies (4, 8, 15, 21). Staphylococcus caprae was found to be unrelated to increased SCC (8). Conversely, Staph. aureus, the main pathogen responsible for clinical mastitis, was detected in this study in persistent subclinical IMI, but was not frequently isolated (5.3%). The low prevalence of Staph. aureus observed in our study was probably due to the selection of goats that were free of clinical mastitis at the start of the study and the use of effective mastitis control practices. Recently, Deinhofer and Pernthaner ( 8 ) studied caprine glands that were naturally infected with staphylococci. Those researchers ( 8 ) found that a significant statistical increase in SCC accompanied IMI caused by Staph. aureus (2950 10 3 cells/ml), Staphyloccocus simulans (900 10 3 cells/ml), and Staph. epidermidis (780 10 3 cells/ml). SCC increases, although not statistically significant, also accompanied IMI caused by Staph. chromogenes (3050 10 3 cells/ml) and Staph. warneri (2430 10 3 cells/ ml). The number of isolates was insufficient for statistical analyses. Those researchers ( 8 ) suggested that these species were potential pathogens for the caprine mammary gland. A comparison of the persistence of caprine staphylococci in the present study with the increase in SCC found by Deinhofer and Pernthaner ( 8 ) indicates some agreement because most of our persistently isolated species were also associated with increased SCC by Dehinhofer and Pernthaner (8). All bacteriological groups yielded false-positive diagnoses, but statistical analysis showed a strong association ( P < 0.001) between isolation of staphylococci and true IMI. From the micrococci isolated, 32.4% yielded false-positive results, which varied by species. All IMI caused by micrococci Staphylococcus hominis, Staphylococcus haemolyticus, and Staphylococcus spp. were false negatives. The CNS group should not be considered a homogeneous group of intramammary pathogens. Further studies are necessary to establish the importance of this group as a pathogen group that affects the caprine mammary gland. Alternatively, the isolation of corynebacteria had a high probability ( P < 0.001) of producing falsepositive diagnoses when a single sample was considered. Corynebacteria are considered to be minor pathogens for the bovine udder and seem to play a similar role in the caprine udder. However, corynebacteria have been isolated in outbreaks of clinical bovine mastitis ( 5 ) as well as in subclinical caprine mastitis (4, 23). In addition, corynebacteria are very sensitive to postmilking teat dipping, and their absence is a good indicator of the benefits of using teat dipping to prevent new IMI by other species of bacteria (20). In the present study, no relationship was found between the isolation of GNB and false-positive diagnoses. The sampling frequency of this study, higher than the usual duration of IMI caused by coliforms, might also have contributed to the low rate of isolation of GNB. However, this low rate of IMI caused by GNB in goats agrees with results from other researchers (19). The high resistance of the goat mammary gland to GNB seems to be related to a higher percentage of intramammary neutrophils in caprine milk (30 to 70%) than in bovine milk (5 to 30%) (19). In addition, the environment surrounding the goat is usually cleaner and drier than the environment surrounding the cow; cows are usually housed, and goats are reared on pasture. Thus, this combination of a heightened natural mammary defense and housing management likely contributed to the low rate of IMI caused by GNB in goats. The data suggest that staphylococci tend to produce persistent IMI, but corynebacteria do not tend to create persistent IMI. The persistency of IMI caused by GNB is unclear from these data, probably because few isolates have been studied. A single milk sample has a predictive positive value of 60% (the probability that a positive gland is truly positive when just one sample is taken). The predictive positive value increased to 67.6% for staphylococci and decreased to 43.8% for GNB. Predictive values are parameters that depend on sensitivity, specificity, and prevalence. Sensitivity and specificity are innate characteristics of a test and do not vary with prevalence (24). To determine whether positive glands are truly positive, two samples must be collected to avoid false-positive results. However, the prevalence detected with a single sample in our study was only 2.5% higher than the prevalence detected with multi-

INTRAMAMMARY PATHOGENS IN GOATS 2819 ple sampling. Because of this, economic aspects, such as traditional costs caused by sampling time, sample transport, and bacteriological analyses, should be considered before dual samples are recommended. When a single sample is used, bacteriological results of isolated groups might be considered (i.e., staphylococci produce true-positive results, and corynebacteria produce false-positive results) to improve the final decision about diagnoses and to avoid an increase in the number of goats considered to be infected, which could be caused by false-positive results. ACKNOWLEDGMENTS The authors thank the goat farmers from ACRIMUR (Asociacion Española de Criadores de la Cabra Murciano-Granadina) for their cooperation. The study was supported by project AGF93-0657- CO2-01 of the Comisión Interministerial de Ciencia y Tecnología. We thank Max J. Paape for critically reading the manuscript. REFERENCES 1 Andrews, R. J., B. J. Kitchen, W. S. Kwee, and F. Duncalfe. 1983. Relationship between individual cow somatic cell counts and the mastitis infection status of the udder. Aust. J. Dairy Technol. 38:71. 2 Bergonier, D., F. Longo, G. Lagriffoul, P. J. Consalvi, and A. Van de Miele. 1996. Frecuence et persistance des Staphylocoques coagulase negative au tarissement et relations avec les numerations cellulaires chez la brebis laitiere. Page 61 in Proc. Int. Symp. Somatic Cells Milk Small Ruminants, Bella, Italy. R. Rubino, ed. Wageningen Press, Wageningen, The Netherlands. 3 Binder, C. 1986. Subclinical mastitis in goats with special reference to Micrococcaceae. Inaugural Diss., Fachbereich Veterinärmedizin, Justus-Liebig-Universität, Giessen, Germany. 4 Contreras, A., J. C. Corrales, and D. Sierra. 1995. Prevalence and aetiology on non-clinical intramammary infection in Murciano-Granadina goats. Small Ruminant Res. 17:71. 5 Counter, D. E. 1981. Outbreak of bovine mastitis associated with Corynebacterium bovis. Vet. Rec. 108:560. 6 Dean, A. G., J. A. Dean, D. Couloumbier, K. A. Brendel, D. C. Smith, A. H. Burton, R. C. Dicker, K. Sullivan, and R. F. Fagan. 1994. Epi Info, Version 6: A Word Processing, Database, and Statistic Program for Epidemiology on Micro-Computers. Ctr. Dis. Control Prev., Atlanta, GA. 7 Deep, H.M.S.S., C. P. Kodikara, and E. A. Wijewantha. 1985. Prevalence and etiology of clinical and subclinical mastitis in goats. Sri Lanka Vet. J. 33:19. 8 Deinhofer, M., and A. Pernthaner. 1995. Staphylococcus spp. as mastitis-related pathogens in goat milk. Vet. Microbiol. 43:161. 9 El-Idrissi, A. H., A. Benkirane, and M. Zardoune. 1994. Studies on subclinical mastitis in caprine herds in Morocco. Rev. Elev. Med. Vet. Pays. Trop. 47:285. 10 Guha, C., A. K. Pramanik, S. K. Misra, and A. K. Benerjee. 1989. Studies on the incidence and diagnosis of subclinical mastitis in goats and in vitro sensitivity of the isolated pathogens. Indian Vet. J. 66:601. 11 Harmon, R. J., R. J. Eberhart, D. E. Jasper, and B. E. Langlois. 1990. Page 33 in Microbiological Procedures for the Diagnosis of Bovine Udder Infection. Natl. Mastitis Counc., Inc., Arlington, VA. 12 Harmon, R. J., and B. E. Langlois. 1995. Mastitis due to coagulase-negative Staphylococcus species. Page 56 in Proc. Natl. Mastitis Counc. Annu. Mtg., Fort Worth, TX. Natl. Mastitis Counc., Inc., Arlington, VA. 13 Honkanen-Buzalski, T., V. Myllys, and S. Pÿorälä. 1994. Bovine clinical mastitis due to coagulase-negative Staphylococci and their susceptibility to antimicrobials. J. Vet. Med. Ser. B 41:344. 14 Hunter, A. C. 1984. Microflora and somatic cell content of goat milk. Vet. Rec. 114:318. 15 Kalogridou-Vassiliadou, D. 1991. Mastitis related pathogens in goat milk. Small Ruminant Res. 4:203. 16 Lerondelle, C., and B. Poutrel. 1984. Characteristics of nonclinical mammary infections of the goat. Ann. Rech. Vet. 15:105. 17 Manser, P. 1986. Prevalence, causes and laboratory diagnosis in subclinical mastitis in goats. Vet. Rec. 118:552. 18 Morant, S. B., F. Dood, and R. P. Natzke. 1988. Consequences of diagnostic errors in mastitis therapy trials. J. Dairy Res. 55: 315. 19 Paape, M. J., and T. Capuco. 1997. Cellular defense mechanisms in the udder and lactation physiology of the goat. J. Anim. Sci. 75:556. 20 Philpot, W. N., and S. C. Nickerson. 1992. Mastitis: Counter Attack. Babson Bros. Co., Naperville, IL. 21 Poutrel, B. 1984. Udder infection of goats by coagulase-negative staphylococci. Vet. Microbiol. 9:131. 22 Sánchez, A., J. C. Corrales, D. Sierra, and A. Contreras. 1994. Relación entre edad y prevalencia de infecciones intramamarias subclínicas en cabras Murciano-Granadinas. Page 177 in Proc. XVII Jornadas Científicas de la Sociedad Española de Ovinotecnia y Caprinotecnia, Albacete. L. Gallego and J. I. Pérez, ed. Ediciones de la Universidad de Castilla-La Mancha, Albacete, Spain. 23 Shouman, M., M. Rezk, M. Ismail, and A. El-Ged. 1986. The role of streptococci and corynebacteria in the subclinical and clinical mastitis in the she-goats and ewes. Assiut Vet. Med. J. 16:341. 24 Thrusfield, M. 1995. Veterinary Epidemiology. 2nd ed. Blackwell Science Ltd., Cambridge, England.